2 Big AI Drops: Sora-2 and Veo 3.1 - RIP Human Creators?

In today’s edition of The AI Catalyst Newsletter:

• 🤯 Sora-2 Just Dropped - and It’s Honestly Terrifyingly Good

• 🤖 This AI Silently Captures My Meetings - Then Does the Follow-Up Work

• ⚔️ Google Veo 3.1 Is Coming — The Quiet Giant Wakes Up

• 🗞 Top 3 AI News You Should Know

  ^{Read time - 4 minutes}

 🤯 Sora-2 Just Dropped - and It’s Honestly Terrifyingly Good

You know that moment when you watch an AI video and think,
“Yeah, that’s cool, but I can still tell it’s AI”?

That moment’s gone.

OpenAI just launched Sora 2 and Sora 2 Pro, and the gap between “generated” and “real” has basically vanished.

The regular Sora 2 is fast and cost-efficient, great for quick B-roll, concept visuals, or short-form storytelling.

You’ll still catch the occasional AI fingerprint here and there, especially with the slightly over-polished skin look of the AI generated people.

But Sora 2 Pro? Whole different story.

It adds audio, better motion physics, and photorealism that crosses into full cinematic territory.

I actually used Sora 2 Pro to generate all the B-roll in my latest video - and it’s unreal how clean it looks 👉 Watch it here !

We’ve now got SORA-2 creating footage that could legitimately pass as a Netflix trailer.
The jump in quality is actually insane.

And here’s why this matters for you 👇

Doors just opened that didn’t even exist last week.
With SORA-2, you can now literally do it all:

🎬 Need B-roll for your content? Create it instantly, no cameras, no budget stress.
📢 Want to run video ads but stuck on cost? Prompt it. Done.
📺 Dreaming of launching your own series? You’ve basically got a full studio on your laptop.
💼 Freelancers? You can now compete with production teams that used to charge six figures.

Right now, the Sora app is invite-only on iOS in the U.S. and Canada.
👉 But if you’re outside those regions, you can already try SORA-2 in a limited form through Freepik, InVideo, and Higgsfield - all live now.

🤖 This AI Silently Captures My Meetings - Then Does the Follow-Up Work

SPONSORED BY RADIANT

You know that feeling when a meeting ends and you think, “Great chat!”
…until you realize you’ve got two hours of follow-up work waiting?

The summaries. The action items. The Slack updates. The “just circling back” emails everyone pretends to love.

This isn’t another meeting bot, like those AI notetakers that crash your Zoom like uninvited guests.
This one never even joins your calls!

Radiant flips the whole thing.
Its an AI assistant for Mac that quietly captures your online meetings and takes care of all the post-meeting work like follow-ups, emails, and updates for you.

It’s honestly been a game changer for me… and this is what my screen looks like every morning 👇

After each meeting, Radiant doesn’t just capture notes - it actually finishes the work. It suggests next steps and executes them automatically:

✅ Drafts follow-up emails (ready to send, written in your tone)
✅ Creates summaries (goodbye, rewatching hour-long recordings)
✅ Updates your tools (Slack, Notion, Gmail - whatever you use)
✅ Schedules your next meetings (calendar tennis = gone)

Setup is ridiculously simple: Download → Connect your calendar → Done in 10 seconds.

Then, it just runs quietly in the background while you focus on real conversations and meaningful work.

👉 Download Radiant on your Mac - totally free while in beta.
(Mac-only for now.)

⚔️ Google Veo 3.1 Is Coming — The Quiet Giant Wakes Up

While everyone’s talking about Wan 2.2 Animate and whispers of Wan 2.5, Google’s quietly preparing its counterpunch - Veo 3.1 - and if the leaks are true, this update could change everything.

According to early reports and insider chatter, here’s what may be coming:
🖼️ Multi-reference input - upload up to 3 angles (front, side, back) to keep characters consistent across scenes.
🌀 Smarter scene expansion - rumors point to mid-shot “outpainting” and reframing for more dynamic composition.
🧩 Improved object editing - refined inpainting and cleanup tools to add or remove details seamlessly.
🎭 Character performance upgrades - more expressive faces, smoother full-body movement, stronger continuity shot-to-shot.
🎥 Cinematic camera moves - pans, tracking, and push shots are reportedly in testing.

If even half of that lands, Veo 3.1 won’t just be playing catch-up - it could become the After Effects of AI video.

And while it hasn’t officially launched yet, Veo 3.1 is already appearing on Higgsfield waitlists, suggesting this platform might be among the first to offer early access once Google flips the switch.

🗞 Top 3 AI News You Should Know

• Replit Launches Connectors - a new way to build apps that talk to Notion, Google, Dropbox, Linear, and more. Watch now

• Mattel Partners With OpenAI to Leverage Sora-2 in Product Design - turning sketches into video concepts using AI video tools. Read more

• Reflection AI Raises $2B - a startup by ex-DeepMind researchers lands a huge round backed by Nvidia, now valued at $8B. Read more